Disrupt or Defend

Learn more about facial vital sign detection: shen.ai & caire.ai

ㅤ

Healthcare has historically lagged in digitalization, creating a significant opportunity for artificial intelligence to jump-start the industry. Host Daniel Kazani sits down with Dr. Lucas Mittelmeier, an investor at Heal Capital, to discuss why the sector's heavy administrative burden makes it a prime target for disruption. They explore the reality of "Shadow AI," where physicians bypass slow hospital IT systems to use tools like ChatGPT for daily tasks. Lucas explains how the industry is splitting into two distinct speeds: highly regulated clinical tools and agile administrative workflows. The conversation also highlights cutting-edge innovations, including facial analysis software that reads vital signs via a camera and vocal biomarkers that detect heart failure.

ㅤ

Guest Bio

Dr. Lucas Mittelmeier is a physician-turned-investor at Heal Capital, a leading European healthtech venture capital firm. With a background bridging clinical medicine, strategy consulting, and startup leadership, he evaluates companies through both medical and business lenses. He is also the author of the Healthtech Off The Record newsletter, where he provides data-driven analysis of industry trends. At Heal Capital, he focuses on sourcing and leading deals from Pre-Seed to Series A.

ㅤ

What We Cover

1. Why the lack of legacy digital infrastructure in healthcare might actually accelerate AI adoption.
2. The phenomenon of "Shadow AI" and why doctors are using consumer tools despite strict hospital regulations.
3. How administrative AI is moving faster than clinical diagnostic tools due to lower regulatory barriers.
4. The potential for "facial parameters" in which video can detect heart rate, blood pressure, and oxygen saturation.
5. Using vocal biomarkers to identify conditions like heart failure by analyzing fluid buildup in the lungs.
6. How typing patterns on a keyboard can serve as early indicators for depression.
7. Why specialized "AI Therapist" startups have struggled to compete with general Large Language Models.
8. The four key moats for healthtech startups: data advantages, network effects, deep customer service, and brand trust.

ㅤ

Resources Mentioned

1. Heal Capital
2. OpenAI (ChatGPT)
3. Anthropic (Claude)
4. Whoop
5. Caire (Healthtech startup)
6. Scale AI

AI Trends in Germany - Presentation (PDF) — Follow along with the data discussed in this episode

ㅤ

Germany currently faces a distinct tension between its technical potential and actual financial commitment to artificial intelligence. While the country ranks high in AI skills and research, private investment stands at just 1.8 billion euros, compared to over 62 billion in the United States. Host Daniel Kazani sits down with Stephan Fricke to examine the reality behind these numbers and what they mean for the German market.

ㅤ

Stephan breaks down the data on Germany's current 45,000 AI specialists and the projected gap of nearly 180,000 by 2032. They discuss why customer contact centers are seeing 88% of implementations and how manufacturing giants like BMW and Siemens are using AI for practical quality assurance. The conversation also covers the critical role of strategic partnerships and outsourcing in bridging the talent shortage that domestic training alone cannot solve.

ㅤ

👤 Guest Bio

Stephan Fricke is the CEO of the Deutscher Outsourcing Verband e.V. (German Outsourcing Association) and the Deutscher Process Automation Verband. Since 2010, he has focused on bridging the gap between German business culture and global innovation hubs. Through industry publications such as the Outsourcing Journal, Stephan shapes the narrative around Global Business Services and advocates for diversifying sourcing destinations to address the talent crisis in the DACH region.

ㅤ

📌 What We Cover

1. The estimated 60 billion euro market volume for AI services in Germany in 2025.
2. Why 88% of German companies implementing AI start with customer contact and chatbots.
3. The massive gap in private AI investment: 1.8 billion euros in Germany versus 62.5 billion in the US.
4. How Germany compares globally in terms of infrastructure, with a notable lack of data centers.
5. The talent crisis: Moving from 45,000 specialists today to a need for 180,000 by 2032.
6. Why Softup and similar partners are becoming essential for companies unable to find local talent.
7. Specific manufacturing use cases for AI: From predictive maintenance at Siemens to quality assurance at BMW.
8. The regulatory hurdles and slow government strategies are affecting European competitiveness.

ㅤ

🔗 Resources Mentioned

1. Deutscher Outsourcing Verband e.V. (German Outsourcing Association)
2. Global AI Index 2024
3. Canva (Marketing tool)
4. Volkswagen
5. BMW Group
6. Siemens
7. Bosch
8. Trumpf

Manufacturing is no longer just about moving atoms. It is shifting toward software-defined automation and fully autonomous systems. Daniel Kazani sits down with Miroslav Kriz, Principal Partner at Momenta, to discuss how AI is reshaping the factory floor. They explore why industrial innovation requires different safety standards than typical software, where a "hallucination" can mean physical danger rather than just bad code.

ㅤ

Miroslav explains the reality of "lights out" factories, where blast furnaces adjust in real time without human input. He also critiques the "tourist syndrome" that European founders face when entering the US market and argues why industrial startups should look to Pittsburgh or Indianapolis rather than Silicon Valley. This conversation covers the journey from simple automation to true autonomy and the specific physics that investors look for before writing a check.

ㅤ

Guest Bio

Miroslav Kriz is a Principal Partner at Momenta, a venture capital firm focused on industrial impact and enterprise technology. He specializes in bridging the gap between legacy industrial companies and modern innovation.

Currently based in Prague after moving from New York, Miroslav works to connect Central and Eastern European technical talent with the US market. He also helps lead initiatives like Gem7 to help startups establish operational beachheads in America.

ㅤ

What We Cover

1. The three core pillars of industrial impact are software-defined automation, robotics, and AI.
2. Why the "move fast and break things" mentality fails in manufacturing, where safety is critical.
3. How virtualization allows agile development on machines with 30-year lifecycles.
4. The emergence of "lights out" factories and autonomous closed-loop systems.
5. Why ROI in industry is defined by speed and waste reduction rather than quality improvements.
6. The "tourist" mistake European founders make when expanding to the US.
7. Why industrial startups often find better success in Detroit or Milwaukee than in Silicon Valley.
8. Using AI in venture capital to validate physics and research trends rather than make deal decisions.

ㅤ

Resources Mentioned

1. Momenta
2. Gem7 (Market entry service)
3. Rockwell Automation
4. Fleet Space
5. Grok
6. Microsoft

This episode was recorded on Dec 10, 2025.

ㅤ

Automation and digitalization were huge topics for decades, but “it’s no longer enough.” Host Daniel Kazani talks with Pascal Faerber, Managing Director, Digital Services Germany at Orange Business, about agentic AI and why it is “fundamentally different” from reactive gen AI. Pascal frames agentic AI as proactive, understanding goals and desired outcomes, breaking them down into steps, executing across multiple systems, evaluating its own output, and learning continuously.

ㅤ

The conversation moves from digital transformation and cloud, including hyperscalers like Azure and Amazon, to a concrete example: a customer success AI agent that scans incoming customer messages across channels, classifies issues, prioritizes urgency, fetches relevant internal knowledge, drafts proposed solutions, triggers actions across systems, and escalates only when human judgment is required. They also talk about AI as a transformation: leadership mindset, processes, and foundations that enable a network of collaborating humans and agents.

ㅤ

Pascal Faerber is Managing Director, Digital Services Germany at Orange Business. He describes Digital Services as “top of the spear” in digital transformation, supporting clients with cloud transformation, data and AI, data platform development, and AI use case development. Pascal also mentions being a lecturer and a business angel, as well as being very active in the tech community.

ㅤ

1. Orange Business Digital Services and “digital transformation,” including cloud transformation and working with hyperscalers like Azure and Amazon
2. A sovereign cloud solution with regulatory requirements and environments operated in Europe by European employees
3. “Agentic AI” as proactive systems that understand goals, breaks them into steps, executes across multiple systems, evaluates output, and learn continuously
4. A customer success AI agent: scanning multichannel messages, classifying and prioritizing issues, pulling contracts, SLAs, documentation, and ticket history, then triggering actions across systems
5. Impact discussed: resolution time brought down from “two, three days” to “15, 30 minutes,” and “60, 65, 70%” reduction in repetitive work
6. AI adoption as a transformation, “mindset first,” and the bottleneck being “permission” rather than technology
7. Clearly defined roles in human and AI collaboration, and AI as “a new colleague.”
8. Moving from pilots to scale: five questions, one high-impact breakthrough, and scaling aggressively with training and change

ㅤ

1. Orange Business
2. Azure
3. Amazon
4. ChatGPT
5. Gemini
6. Salesforce
7. Nvidia
8. Microsoft
9. IBM
10. Satya Nadella

AI has shifted from a buzzword to a genuine, intelligent assistant. Host Daniel Kazani talks with Dajana Stojchevska, a senior software engineer at Google in Munich, about how AI is embedded in day-to-day engineering work, not about replacing engineers, but about boosting productivity and removing friction so teams can focus on the cool stuff.

ㅤ

Creation, collaboration, and knowledge management show up everywhere, from intelligent code completion inside the integrated development environments, to AI-assisted code reviews, to meeting notes that summarize transcripts, highlight key decisions, and list action items with owners. The conversation also remains grounded in the challenges: the hallucination trap, prompt injection, indirect injection hidden in external data such as a PDF, and strict discipline around data privacy. Looking ahead, the focus turns to autonomous agents, massive context windows, proactive analysis, and the evolving role of the software engineer as architect and orchestrator.

Dajana Stojchevska is a senior software engineer at Google in Munich. She graduated with a degree in Scopia from the Faculty of Computer Science and Engineering, with elective subjects in software engineering. She completed a few internships, including in Python, and her first role was as a Java developer focused on full-stack web development with Java and Angular. She also worked as a laboratory teaching assistant, helping students with exercises. After about two years, she moved to Germany for the Google offer.

ㅤ

1. AI inside the editor as a proactive teammate, intelligent code completion, prompts for snippets, and boilerplate
2. Code reviews with AI, drafting descriptions, style and standards fixes, and automated fixes from static analysis errors
3. AI-generated suggested code edits from teammate feedback, plus reviewer support with links to documentation
4. Meeting notes that summarize transcripts, highlight key decisions, and list action items with owners
5. Generating architecture diagrams from text, plus document analysis and “interviewing the document”
6. Correctness and the hallucination trap, treating AI like a junior engineer who needs supervision
7. Security risks, direct prompt injection, indirect injection, and why even a PDF can be a hostile input
8. Data privacy and strict guidelines on what data can go into which tools, plus internal AI chatbot support
9. Staying up to date with internal channels, newsletters, tech talks, hands-on daily practice, and peer community
10. The next five years: autonomous agents, bigger context windows, proactive help, and engineers as architects and orchestrators

ㅤ

1. Gemini
2. Gemini Three
3. Anti-Gravity
4. Fathom
5. Notebook LM
6. Integrated development environments
7. Static analysis tools
8. Internal AI chatbot

Being discoverable on ChatGPT has become a board-level problem for most companies. Host Daniel Kazani talks with Vlad Shvets about Qvery, an AI agent software that helps brands measure and grow their visibility and share of voice on the AI search engines. Measuring brand visibility is a new challenge companies need to address, requiring new tools and software. The way we search Google is very different from how we use ChatGPT, which provides comprehensive recommendations personalized to your conversation history, language, and location. The quest starts when a CEO goes to chat, asks a question related to their brand, and the brand does not appear. The marketing team begins assessing how to measure this and how to make the brand discoverable. Vlad breaks down three pillars: your own website presence, mentions on external websites, and user-generated content, and why Reddit is the most critical website these days.

ㅤ

Vlad Shvets is a marketing expert, serial entrepreneur, advisor, and founder of Qvery. Qvery is an AI agent software that helps brands measure and grow their visibility and share of voice on the AI search engines. Vlad shares that Qvery started as a consultancy, then evolved into a way to measure visibility on chat GT, including personalized results. He describes a vision of the future of the web as agentic, with people increasingly relying on AI agents to do tasks for them.

ㅤ

1. When a CEO goes to ChatGPT, asks a question related to their brand, and the brand does not pop up
2. Google gives you links, and ChatGPT gives you complete recommendations, with personalization
3. A case study for services, a separate domain, a single-page website, and leads from ChatGPT and Google AI overviews
4. Optimizing for specific granular use cases, capturing high-intent requests, and vanity metrics
5. Three pillars: your own website presence, mentions on external websites, and user-generated content, Reddit in particular
6. FAQ schema and schema data as fast food for chat gt to fetch and understand
7. Logged in state of personas, citations list, outreach, and getting a product mentioned where it matters
8. CloudFlare blocking agents, browser manipulation tech, AI agent regulation, and a passport program
9. Google AI mode is becoming a default way to search, and it's what happens overnight for companies and businesses

ㅤ

1. ChatGPT
2. Google AI mode
3. Gemini 3 model
4. OpenAI
5. Reddit
6. Perplexity
7. CloudFlare
8. Browserbase
9. Browser manipulation tech
10. Schema, FAQ schema
11. API
12. Slack
13. Wikipedia
14. Patreon

We are living through the most significant platform shift since the Internet. Host Daniel Kazani talks with guest Shefqet Avdullau, an angel investor, advisor, and speaker focused on growth-stage B2B SaaS, FinTech, ad tech, and health tech. The conversation starts with a story that moves from coding to multiple ventures to a meaningful exit, then into investing with a mentor who gave a head start on due diligence, pitfalls, and strategies. The weight falls on the team because the idea you start with does not necessarily mean you will end with it, and a good team can turn a bad idea into a great one. Then: AI and defensibility, wrappers, data-loop strategy, fine-tuning, and what happens if OpenAI or Gemini releases a new update tomorrow. Health tech and biotech, drug discovery, and turning biology into an engineering problem.

ㅤ

Shefqet Avdullau is an angel investor, advisor, and speaker. He invests in serial founders across the US and UK, focusing on B2B SaaS, FinTech, and ad tech at all stages, and on health tech specifically at the growth stage. His foundation is in tech; he worked in that field for about 13 years, started multiple ventures with some small exits, and then had one meaningful exit. In about four years, he has done about 16 investments and has had two exits.

ㅤ

• From coding, to multiple ventures, to a meaningful exit, to investing and joining a group of investors
• A mentor with private equity experience, due diligence, pitfalls in investing, and strategies to follow
• Why serial founders come with a map, with a playbook, and go straight to finding product market fit
• Scars, lessons, when things get tough, and why failure can be something you prefer
• Idea versus team, pivots, and why the team can turn a bad idea into a great idea
• Two founders or more, complementary skillset, product, and sales, and a third on operations
• Founder problem fit, domain experience, network, and solving an actual problem, not just for money
• AI wrappers versus defensibility, data loop strategy, fine-tuning, and “would this company die” after a new update
• Where AI is disrupting, health tech and biotech, drug discovery, simulating millions of interactions digitally, and FinTech underwriting with unstructured data
• Using AI for competitor analysis, risk analysis, and alternative potential revenue streams, and “it hallucinates a lot”
• A contrarian investment choice, two serial founders, employee disengagement, productivity, and invisible frictions

ㅤ

• Open AI
• Gemini
• Figma
• Nvidia
• LinkedIn
• lovable

A wave of excitement and activity around AI is hitting management consulting, and many leaders are asking the same questions. What is this AI thing? What do I do with it? And what does it really mean for my business? In this episode, host Daniel Kazani talks with Matthew Murphy, a partner at AMEND Consulting in Cincinnati, about how technology and AI now sit at the forefront of almost every new client conversation.

ㅤ

They explore why AI discussions often uncover missing fundamentals in process, technology, and data management, and why the biggest wins today show up in highly manual, tedious, document-intensive, and task-intensive work. From order entry and AR and AP automation to image recognition in retail stores and AI-supported assessments in consulting, they share concrete examples of AI agents working in a human-in-the-loop way. The conversation then moves to augmentation versus role replacement, departments that cannot fill roles, the new workforce entering the market, and how AI is reshaping the core business model of professional services and long-term client relationships.

ㅤ

Matthew Murphy is a partner at AMEND Consulting, a management consulting firm based in Cincinnati. He has spent over a decade driving transformation for mid-market and large enterprises, working across people, process, and metrics. He also ran and led a software business at AMEND for four years. Now he helps clients in the age of AI and automation, helping them leverage technology and grow smarter across operations, analytics, and automation.

ㅤ

• Why technology and AI are at the forefront of almost every conversation with new or existing clients, from enthusiasm to the fear of falling behind
• How AI projects often start with automating a particular process, but lead to missing foundations in process, technology, and data management
• Real-world automation in order entry, where customer service and inside sales teams spend most of the day keying in orders, and AI agents can now do a bulk of the work
• Accounting and finance use cases like AR and AP automation, financial close, and reconciliation, and how broad workflow and AI automation tools can be applied across many process areas
• Image recognition in retail, where store audits used to mean hundreds of pictures per store and manual review, and an AI agent now sifts through images with a high degree of accuracy and improves the quality of life for the team
• How AMEND assessments have changed from heavy note-taking and weeks of brute force compilation to AI agents that process notes, meeting recordings, and GoPro footage and produce first-pass gap and theme compilations in hours or days
• AI as a way to capture and expose tribal knowledge from hundreds or thousands of work instructions, helping a newer workforce get up to speed more quickly, instead of hunting through file repositories or an LMS
• The reality of augmentation versus role replacement, from overworked teams doing two people’s worth of work to departments that choose not to fill open roles because AI enables the same team to do more
• Why AMEND is raising the watermark for technical competency for new hires, partnering with universities, and still investing in junior talent even as some larger firms cut hiring targets
• How AI challenges the traditional dollars for hours model in professional services, pushes firms toward value-based pricing, and increases the importance of being a trusted advisor focused on long-term relationships and business impact

ㅤ

• AMEND Consulting
• RPA tools for process automation
• AI agents for workflow and document-intensive tasks
• LMS and file repositories as traditional knowledge bases
• GoPros for process shadowing and manufacturing floor observations
• Nvidia is an example of developers experimenting with new tools every day

AI is changing how software is built and how companies run day-to-day at Softup. Host Daniel Kazani sits down with co-founder and CTO Kristi Kristo to walk through concrete examples of how AI touches almost every part of the business. They talk about perfect developer profiles, automated estimations, and an internal Weekly Digest that helps decision makers spot opportunities and problems faster in a distributed team.

ㅤ

On the technical side, they share how Cursor and AI agents act as a co-developer, how some features ship with almost zero manual code, and why quality can even improve when context is structured well for the LLM. Kristi explains his bold goal of reaching zero manual code, why coding is only one part of software engineering, and how the role of the developer is moving closer to product, business context, and orchestration. They close with what AI transformation looks like for founders and SMEs today and why AI Labs at Softup experiments with the latest tools so customers can benefit from real, applied AI.

ㅤ

Kristi Kristo is the co-founder and CTO of Softup and Managing Director at Softup Technologies GmbH. His focus is on AI Engineers, AI Agents, and MVPs that scale as AI transforms how companies build and deliver products. Kristi describes his work with a simple line: AI Agents will transform every business - including yours. We build the systems that make it happen. At Softup Technologies, he leads teams that use advanced AI tools and workflows to deliver software faster while keeping a strong focus on real business problems.

ㅤ

• How Softup uses AI on the business side to create perfect developer profiles, cut grammar mistakes, and avoid missing relevant experience when sending CVs to customers.
• Why automated estimations with AI remove 80 to 90 percent of the brain capacity and effort from the team, and how this turns two or three estimations per day into a streamlined process founders can rely on.
• The Weekly Digest automation workflow that collects what everyone did, what they will do next, and helps decision makers spot opportunities, problems, and availability across a distributed team.
• How developers at Softup use Cursor as an AI co developer, spin up multiple AI agents in parallel, and sometimes ship features and modules while writing almost zero manual code.
• Why Kristi believes writing manual code will go close to zero, why coding is only 30 to 70 percent of a developer’s time, and how orchestration, architecture, testing, and understanding business context become even more important.
• How the day to day of a developer has changed since the LLM world, with ChatGPT, Cursor, codex, sonnet, cloud code, and Code XCLI always open as part of the normal workflow.
• Why Kristi thinks newcomers may not always need to know code deeply if AI agents for testing, security, and cloud give a thumbs up, and why shipping and orchestration skills matter more over time.
• What Kristi and Daniel Kazani see in the market: a two year lag between the first OpenAI release and real pressure from CEOs, boards, and investors to invest in AI across customer support, finance, sales, guest experience, PropTech, FinTech, and more.
• How Kristi splits AI work into automation workflows and AI agents, why processes need to change in an AI first world, and why AI transformation in marketing, sales, operations, HR, engineering, and customer support is a long term journey rather than a quick project.
• The idea behind AI Labs at Softup, from building experiments with MCP and local deployment to testing OpenAI commerce and ChatGPT apps, including the first ChatGPT app for the hospitality industry and use cases like helping FinTech merchants sell inside ChatGPT.

AI impresses Xhoni Shollaj almost every day, from protein folding and the idea of a virtual cell to autonomous driving and robotics. In this conversation, host Daniel Kazani follows Xhoni’s journey from business studies and data roles at PwC and EY in Albania and Bulgaria to a Master of Science at the National University of Singapore and his current work as a senior AI solutions architect at NVIDIA.

ㅤ

The discussion moves from early natural language processing and computer vision projects, document reduction and summarization tools, to building and maintaining large scale language model applications. Xhoni shares how staying in touch with GitHub trending projects, arXiv style paper feeds, and the open source community shaped his path. Founders, CTOs and decision makers hear concrete talk on AI experiments versus production systems, scalability, security, hallucinations, golden datasets, vibe coding, tools like Cursor, ChatGPT and Gemini, and why contributing to open source with teams at NVIDIA, Google and others can be a powerful way to stand out.

ㅤ

Xhoni Shollaj is a Senior AI Solutions Engineer at NVIDIA, specializing in developing and deploying large language model architectures. He started with business, moved into computer science, and began his AI journey in research and development teams at PwC and EY in Albania and Bulgaria, building machine learning based applications and automation solutions. Xhoni then joined the National University of Singapore, working on internal automation tools and research support for patents and papers, before moving into his current role at NVIDIA in Asia.

ㅤ

• How Xhoni moved from business studies and data roles at PwC and EY in the Balkans to a Master of Science at the National University of Singapore and into AI solutions work at NVIDIA.
• Why he chose Singapore for its faculty, research direction and blend of cultures, and how being location agnostic helped him follow the strongest data science programs.
• The habits he sees as most useful for people who want to succeed in AI, including staying in touch with the latest technologies, GitHub trending, arxiv style feeds, open source projects and strong news sources.
• Areas where AI feels most disruptive today, from protein folding and the path toward a virtual cell for drug discovery and disease treatment to space exploration, SpaceX and ideas like space data centers.
• How to distinguish an AI experiment or small POC from a production system, with concrete points on autoscaling, multi cloud and multi zone backups, security pipelines, identity and access management, encryption, multilingual behavior, hallucination tracking and observability.
• Approaches to accuracy and hallucinations, including well built RAG pipelines, choosing the right benchmarks and metrics, literature reviews, leaderboards, human in the loop evaluation and tracing problems back to data sources or model behavior.
• The reality of vibe coding for non technical founders, why it is a net positive and equalizer, and how to combine fast POCs with later help from experienced engineers on scaling, security and edge scenarios.
• Tools and workflows Xhoni personally uses, such as Cursor with Claude 4.5 Sonnet, ChatGPT and Gemini for brainstorming, creating plans, testing ideas and even asking models to make fun of an idea to expose weak points.
• The most common challenge companies face when integrating AI into their business, why a golden dataset and clean, validated, well reviewed data can make or break a project, and how synthetic data and diverse scenarios help test chat bot performance.
• Why AI systems in sectors like hospitality need clean booking and address data, strong formatting, and synthetic test scenarios with mixed languages, toxic and non toxic inputs, and special characters to evaluate real world behavior.
• Thoughts on interpretability and mechanistic interpretability, the black box nature of transformer layers today, and why being able to trace reasoning in sensitive areas like healthcare or drug simulation matters.
• How different large models like OpenAI, Claude, Gemini, DeepSeek and NVIDIA models feel to users because of data sources such as Reddit and prompt level instructions, leading to different levels of confidence, politeness and directness.
• What a billion plus dollar AI data center collaboration between NVIDIA and Deutsche Telekom in Germany might mean for telecom, communication, research and startup ecosystems in Munich, Germany and across Europe.
• Final advice for students, graduates and people struggling in the current environment, including a clear call to contribute in their free time to open source NVIDIA and Google projects, build relationships, learn in public and stand out through real work.

ㅤ

• NVIDIA
• PwC
• EY
• National University of Singapore
• AWS
• Azure
• Deutsche Telekom
• SpaceX
• GitHub trending
• Arxiv style feeds and “alpha Arxiv sanity”
• Cursor
• Claude 4.5 Sonnet
• ChatGPT
• Gemini
• DeepSeek
• Tinker
• OpenAI
• Anthropic
• Google
• Reddit